Craig Park, FSMPS
Strategist · Author · AEC Advisory
When AI takes on the first pass, AEC firms need to decide where emerging professionals will learn to weigh evidence, make tradeoffs, and stand behind a recommendation.
Picture a young designer preparing for a meeting about a campus renovation. She gives an AI assistant the project team's notes and asks it to compare three phasing options. In minutes, she has a readable matrix and a recommendation. The principal looks it over, changes a few words, and sends it to the client.
Then the facilities director asks, "What happens if we can't close that entrance in August?"
The matrix treats the closure as a given. The designer can find the line in the notes where it came up. She's less sure who agreed to it, whether another route would work, or how much the recommendation depends on it.
That question from the facilities director is where the education begins. The trouble is that on too many projects, it could also be the first time the designer has been asked to defend the reasoning.
The learning hidden inside the work
In architecture, engineering, and construction, much of what we call judgment develops while doing ordinary tasks. A junior professional draws an option, follows its consequences into the next discipline, and discovers why a line that looked harmless on one sheet creates a problem on another. A senior colleague asks a question. The junior goes back to the drawing, this time knowing what to look for.
Even an awkward first draft has something to teach. It reveals what its author noticed, what they assumed, and where they got stuck. An experienced reviewer can work with that. A polished AI draft can leave those questions harder to see, especially when the review starts with edits to the wording.
There's no particular virtue in making someone spend a morning formatting a matrix. There is value in having them decide which criteria belong in it, which evidence they trust, and what would change their recommendation. As firms shorten production tasks, those decisions need a place in the workday.
Give the emerging professional a decision to make
Before opening the AI draft, ask the person responsible for the work to write down three things: What decision does the client need to make? What fact or assumption could change my recommendation? What would I verify before putting my name on it?
On the campus project, that might take five minutes. The designer might write that the university needs to choose a sequence that keeps a usable public entrance open through orientation. She could flag the August closure as an assumption and ask campus operations who owns that decision.
Now the team can use AI to compare alternatives against a problem the designer has framed. It can even ask the tool to make the strongest case for the option the team initially dismissed. In review, ask the designer to explain which argument held up and what still needs checking. That gives the principal something specific to coach and gives the designer practice revising a recommendation in light of evidence.
Let people hear how other disciplines think
The campus entrance looks different to a designer, a contractor, and the person who runs the building. The designer sees circulation. The contractor sees sequencing and temporary barriers. The operator knows where orientation volunteers actually direct people. Put those perspectives together and an option that scored well on paper may lose its advantage.
Invite an emerging professional to lead part of that conversation. Ask them to explain how the recommendation changed after hearing from the other disciplines and the client. A senior colleague can make their own judgment visible in return. A sentence like that teaches someone what to listen for on the next project.
This is also how a firm builds a community of practice. People learn whose experience to seek out, how to disagree without stalling a decision, and how to pass along what they learned. Those relationships matter when a project needs an answer that no single discipline can supply.
Put apprenticeship into the project plan
Pick one recurring task where AI is already saving time. It could be comparing design options, summarizing a coordination issue, or drafting a client recommendation. For the next project, give an emerging professional ownership of one bounded decision. Put a short review on the schedule before the recommendation goes out. Ask the reviewer to explain one consequential edit in terms of its effect on the client, the project, or the people who will use the building.
At the next milestone, give that person a similar decision with less direction. See what they catch on their own. If they notice an assumption sooner, bring in the right colleague earlier, or explain a tradeoff more clearly to the client, you have evidence the firm is developing judgment along with improving production.
When the facilities director asks about that entrance in August, the answer may still be, "I need to check." What you want is a professional who can tell her exactly what they need to check, who they'll ask, and how the answer could change the plan. That's a skill worth making time to teach.
More from CatalystFactor
A Catalyst Factor Essay
Strategy Is No Longer a Plan. It Is a Practice.